OpenAI doubles support for Lenfest AI fellowship in local newsrooms

A larger program after two years

The Lenfest AI Collaborative and Fellowship Program places full-time AI technologists into operating news organizations to drive business innovation in local journalism. It launched in 2024 with AI engineering fellows in 11 major American news organizations, and the Lenfest Institute describes it as the largest AI fellowship program in American journalism.

In the next phase, the Institute will invite a new group of news organizations to join the cohort. It says the program will continue to embed AI fellows within participating newsrooms while broadening the range of organizations involved.

Fellows work with newsroom leaders, reporters, product teams, revenue staff and executives to identify organization-specific problems and build practical AI engineering solutions, along with what the announcement calls safe, transparent and ethical policies. Several fellows are expected to stay on as full-time employees at multiple organizations.

Examples from the pilot

The announcement lists projects across the 11 organizations, covering audience engagement, news product development, investigative research, advertising and reader revenue operations, and community reach.

• The Philadelphia Inquirer: Dewey, a tool that helps journalists search decades of archived reporting, and Scrape, a monitoring tool that turned a roughly 15-hour-per-week reporting task into a daily digest of potential story leads from around the newspaper's region.
• Chicago Public Media: AI-assisted translation workflows that allowed time-sensitive Spanish-language coverage to be published in a fraction of the time previously required, and the transcription of audio archives that had been in storage for decades.
• Other organizations: projects on advertising prospecting, donor modeling, audience personalization, subscription growth, public-meeting monitoring, print-to-digital production transition and operational efficiency.

Fellows also shared code and ideas between newsrooms. The Baltimore Banner, for example, drew on work from The Philadelphia Inquirer's team while building its own news discovery tools.

Lessons: trust, problems first, collaboration

The organizers name three main lessons. First, trust matters more than technology. Fellows were hired locally and embedded for two years, and they acted as team members rather than outside consultants. According to the announcement, adoption grew as journalists came to see them that way.

Second, projects worked best when they started with a clearly defined challenge rather than a predetermined technical solution. Successful ones involved close work with users, small-scale experimentation and ongoing evaluation.

Third, collaboration accelerates innovation, because one newsroom's breakthrough often became another's starting point.

Lenfest Institute Executive Director and CEO Jim Friedlich said the most valuable lesson of the first two years is that "responsible AI adoption is less about the technology and more about trust."

What comes next

The Institute plans to expand its technical capabilities so it can turn the strongest fellowship projects into reusable tools, frameworks, plugins, implementation guides, playbooks and technical resources. These are intended to benefit hundreds of news organizations. One stated goal is to make the cross-newsroom learning seen in the pilot easier to sustain and scale, so that a tool or workflow built in one organization can become shared infrastructure for the wider news field.

Tom Rubin, OpenAI Chief of Intellectual Property and Content, said it has been "wonderful to see so many practical examples of journalists using AI as a tool to help deliver quality local news." He added that OpenAI wants to expand the partnership to support more journalists.

For local publishers, the announcement points to a wider cohort and to shared resources that could be used beyond the fellowship itself. The source does not give details on how many new organizations will join or when.